Saturday, January 16, 2010
Google's Downfall - Quality Control
Well, to be more precise, the lack of human intervention. This is something Google has been very proud of - no humans meddling with their search results. The legendary algorithm makes sure that only quality sites appear in the search results.
Oops! Quality is diminishing. I'm seeing more and more spam in the search results. And worse still, pseudo-content. This is what is happening:
Dozens of automated processes are bombarding Google with automated content, and Google is not keeping up. As soon as they tackle the current batch, the next batch of automations arrive and there's even more of them
But worse still - humans are churning out the stuff. It doesn't matter than someone in India writes it for $15, or an expert in the field writes a worthy article - Google's algorithm cannot tell the difference. And unfortunately, some humans will genuinely link to it, because they can't tell the difference!
And in a few years, computer will do the entire process - discover what people are searching for, and create content to suit. And the end user will use software to find what fits their interests... At some stage, we will need humans again.
Oops! Quality is diminishing. I'm seeing more and more spam in the search results. And worse still, pseudo-content. This is what is happening:
Dozens of automated processes are bombarding Google with automated content, and Google is not keeping up. As soon as they tackle the current batch, the next batch of automations arrive and there's even more of them
But worse still - humans are churning out the stuff. It doesn't matter than someone in India writes it for $15, or an expert in the field writes a worthy article - Google's algorithm cannot tell the difference. And unfortunately, some humans will genuinely link to it, because they can't tell the difference!
...companies like Demand Media and Answers.com which create thousands of pieces of content per day and are making a big impact on the Web. Both of those two companies are now firmly inside the top 20 Web properties in the U.S., on a par with the likes of Apple and AOL.Demand are churning out 2,000 articles and videos per day.
The average writer earns $15 per article for pieces that top out at a few hundred words, and the average filmmaker about $20 per clip, paid weekly via PayPal. Demand also offers revenue sharing on some articles, though it can take months to reach even $15 in such payments. Other freelancers sign up for the chance to copyedit ($2.50 an article), fact-check ($1 an article), approve the quality of a film (25 to 50 cents a video), transcribe ($1 to $2 per video), or offer up their expertise to be quoted or filmed (free). Title proofers get 8 cents a headline. From Wired.The quality of Demand's articles and videos is passable. But the next company to do the same will have less quality, and I suspect Google will fall for it.
And in a few years, computer will do the entire process - discover what people are searching for, and create content to suit. And the end user will use software to find what fits their interests... At some stage, we will need humans again.
Labels: demand media, google
Saturday, May 30, 2009
Status - The Key to the Survival of Phone Companies
It's way to late for me or anyone else to come up with this idea. It could only be of use if every phone company in the world decided on a standard and implemented it within a year. Sounds impossible.
The idea: just like you can set a status in IM, you should be able to do it on your phone. For example, if I am in the cinema, I don't want to take calls period. And I don't want people to leave messages. But if I could set my status to unavailable, and my status message to "in cinema until 8:15", then people would understand, and call me back after the movie finishes.
Specs:
Status options of Available, Unavailable & Busy
Status Messages of a length up to 130 chars, leaving 14 for the status type
Options of who gets to see your status, or your status and your message
The ability to change the status dependant on who is calling
That's it - all that is needed. The rest can be handled by software on the phones.
Phones can then do things like "if Bob calls, and my status message is "asleep", let the phone ring.
In doing so, people will retain their phone services, rather than switching to VOIP, IM etc. Otherwise, bye-bye phone companies, hello ISPs.
The idea: just like you can set a status in IM, you should be able to do it on your phone. For example, if I am in the cinema, I don't want to take calls period. And I don't want people to leave messages. But if I could set my status to unavailable, and my status message to "in cinema until 8:15", then people would understand, and call me back after the movie finishes.
Specs:
Status options of Available, Unavailable & Busy
Status Messages of a length up to 130 chars, leaving 14 for the status type
Options of who gets to see your status, or your status and your message
The ability to change the status dependant on who is calling
That's it - all that is needed. The rest can be handled by software on the phones.
Phones can then do things like "if Bob calls, and my status message is "asleep", let the phone ring.
In doing so, people will retain their phone services, rather than switching to VOIP, IM etc. Otherwise, bye-bye phone companies, hello ISPs.
Labels: mobile status
Sunday, May 17, 2009
The #1 Problem with the Rise of Video
As one becomes more experienced in surfing the web, one learns to quickly judge whether a web page contains credible, useful information, or not. The clues are not necessarily obvious, but deep-down you've learned to spot them:
- more ads than content
- brevity
- ads not related to content
- ads for gambling, credit cards, adult dating
- no sign of an author's name
- poor writing quality in the first paragraph
- lack of related imagery
- domain name with hyphens
- domain name using generic keywords
- lack of links to related content
- links within the content that lead to ads (commonly with a double-underline)
...and that's just off the top of my head. Most of time I can choose whether to read the content of a page, or not, in less than a second.
With video, most of the time all I have to judge it by is the title and the opening frame. That is next to nothing. Given how easy it is put together a professional looking, cut and paste video these days (I wouldn't have a clue how, but given how many appear on YouTube each day, it must be easy), a good portion of videos you come across are a waste of time. If you only ever visit the BBC you won't have a problem, but the more you stray from super-trustworthy sites, the lower the quality of the embedded videos you are likely to encounter.
Ongoing, people will need to learn how to decide which videos are worth watching. This will be a serious skill, because you cannot judge a video on the first second. Some don't even really get going until 10, 20 even 30 seconds from the start. And even then, if it takes them a while to get to the thing you desire to see, you might stick around, waiting...
These indicators of quality might help:
- personal knowledge of the quality of the site
- surrounding article - the more supporting words the better
- is it something that needs to be on video rather than words?
Regarding the last factor - UFO footage would be a definite, so would a celebrity interview or a plane crash. Or a demonstration of a product - I love video reviews of tech gear. But if it is a concept or theory, give me an article over a PowerPoint presentation any day.
In the future, I can see two things happening:
- a rise in bait and switch videos
- a rise in rating services
A bait and switch video is where you start watching one thing, and slowly but surely the content changes to something else. Perhaps a video on fixing credit problems that morphs into a pitch for a pyramid scheme. We know from infomercials on TV that people will continue to watch something once they have given it their initial attention.
Rating services could be in the form of a toolbar with thumbs up, thumbs down buttons. Or a paid service. Plenty of commercial website have badges in the footer of the page to indicate how trustworthy they are. Why not take this a step further and have a video certified to be of a high standard?
Or, we might just find that, unless it is from a site they trust, people will just stop watching videos.
- more ads than content
- brevity
- ads not related to content
- ads for gambling, credit cards, adult dating
- no sign of an author's name
- poor writing quality in the first paragraph
- lack of related imagery
- domain name with hyphens
- domain name using generic keywords
- lack of links to related content
- links within the content that lead to ads (commonly with a double-underline)
...and that's just off the top of my head. Most of time I can choose whether to read the content of a page, or not, in less than a second.
With video, most of the time all I have to judge it by is the title and the opening frame. That is next to nothing. Given how easy it is put together a professional looking, cut and paste video these days (I wouldn't have a clue how, but given how many appear on YouTube each day, it must be easy), a good portion of videos you come across are a waste of time. If you only ever visit the BBC you won't have a problem, but the more you stray from super-trustworthy sites, the lower the quality of the embedded videos you are likely to encounter.
Ongoing, people will need to learn how to decide which videos are worth watching. This will be a serious skill, because you cannot judge a video on the first second. Some don't even really get going until 10, 20 even 30 seconds from the start. And even then, if it takes them a while to get to the thing you desire to see, you might stick around, waiting...
These indicators of quality might help:
- personal knowledge of the quality of the site
- surrounding article - the more supporting words the better
- is it something that needs to be on video rather than words?
Regarding the last factor - UFO footage would be a definite, so would a celebrity interview or a plane crash. Or a demonstration of a product - I love video reviews of tech gear. But if it is a concept or theory, give me an article over a PowerPoint presentation any day.
In the future, I can see two things happening:
- a rise in bait and switch videos
- a rise in rating services
A bait and switch video is where you start watching one thing, and slowly but surely the content changes to something else. Perhaps a video on fixing credit problems that morphs into a pitch for a pyramid scheme. We know from infomercials on TV that people will continue to watch something once they have given it their initial attention.
Rating services could be in the form of a toolbar with thumbs up, thumbs down buttons. Or a paid service. Plenty of commercial website have badges in the footer of the page to indicate how trustworthy they are. Why not take this a step further and have a video certified to be of a high standard?
Or, we might just find that, unless it is from a site they trust, people will just stop watching videos.
Labels: problem, spam, video search
Monday, May 11, 2009
This Will Be Big - Online Game Show
Microsoft is debuting an interactive version of "1 vs. 100," an Endemol-created game show in which one person combats 100 others for a sizable cash prize."1 vs. 100" marks the first offering for Xbox Live Primetime. The show will air in North America over the summer and fall, MediaWeek writes, and as many as 200,000 Xbox Live Gold members are expected to play either the live versions, which will air Fridays and Saturdays, or quick-play versions throughout the week.
I'm not predicting that this first iteration will be a hit, but the concept in general. Everyone that watches a game show wishes they were playing it. Online PCs and consoles are very capable of running muti-player games, even massive ones, so such contests should be easy to host, and popular. Early adopters clever at finding ways to cheat could intially make some good $$
Labels: 1 vs. 100
Wednesday, April 22, 2009
Twitter Tiny URL service
There has been lots of talk lately about Twitter, TinyURL, and other URL shortening services. To me the solution is obvious....
- Twitter develops their own URL shortening service. Shouldn't be hard, such scripts can be purchased for $10...
- Twitter adjusts their software, so that when they check the length of any tweet, the first URL is counted as just 7 characters
- URLs embedded in tweets just appear like this uGT5ER7 - with the u denoting it is a Twitter short URL
- If Twitter also launched a browser toolbar, with a big button for Tweet This Page, which automatically embedded a short URL for the page you are visiting, then the process is complete
Labels: twitter
Tuesday, April 14, 2009
19% of Google sessions now come from Facebook
So the article says. Extremely misleading, and the author then extrapolates this to suggest that "Facebook could kill Google".
Here's what is said:
Here's what is said:
Ross also illustrates how important Facebook has become to Google as a traffic source. Fully 19% of Google sessions now come from Facebook, up from 9% a year ago. At the very least, this will likely give Facebook the leverage to negotiate a sweet referral deal at some point.Here's the reality:
- Facebook does not link to Google (unless some places a link, or a YouTube video on their page)
- The traffic comes from people finishing whatever they were doing at Facebook, and typing google.com directly into the browser
- This result is directly related to the amount of time people spend on such a popular site. It would be the same for other popular sites
- Google would not give Facebook anything for these referrals, because they are purely related to the popularity of Google; Facebook isn't doing anything to help
Wednesday, April 1, 2009
Farewell Wikia Search
Well, a few people will be saying farewell, but most of us didn't care for or use Wikia Search. With only 10,000 unique visitors per month it just didn't have enough users to work. About the only good it has done is prompting Google to enable users to personalise results.
Wikia Search had the right idea, but the wrong methodology. The search engine that knocks Google off their perch will be mostly automated, but augmented by paid staff - experts that tweak results and remove spam. For example, if an algorithm can show me websites that it thinks are spam, I could probably verify their spammy-ness at a rate of 20 per minute. The algorithm can then use this knowledge to delete in bulkfrom the index similar sites made by the same people... I could go on. Google, hire me! (again).
Wikia Search had the right idea, but the wrong methodology. The search engine that knocks Google off their perch will be mostly automated, but augmented by paid staff - experts that tweak results and remove spam. For example, if an algorithm can show me websites that it thinks are spam, I could probably verify their spammy-ness at a rate of 20 per minute. The algorithm can then use this knowledge to delete in bulkfrom the index similar sites made by the same people... I could go on. Google, hire me! (again).
Labels: wikia search
Subscribe to Posts [Atom]
